IDSCA, an intelligent system, is developed for the direction selection of controller's action in a multiloop control system. In the design of a controller, the selections of both the valve type and the controller's action direction are important tasks, which directly affects the operation and safety of production. Traditional design can hardly solve the problem. Programmed in OPS5, IDSCA can perform the heuristic inference reasoning and make the intelligent decision. A significant result from IDSCA is the fact that a new design criterion is developed, which may complement the knowledge of controller design technique. The other important investigation is that the Adaptive Feedback Testing System (AFTS) is developed to provide the high reliability of the design results. These two investigations indicate that the development of'intelligent systems can stimulate and help the development of both A1 and related prototype problems. Moreover, IDSCA has some additional important features: its knowledge base can be modified and new production rules can be created in the running process to solve special problems; and the hierarchy of meta-level control strategy provides the means to manage the knowledge base of IDSCA efficiently. In this article, the principle of building intelligent systems is discussed. As an example, the cascade control system of a polymerizer is applied to illustrate the use of IDSCA.